Trustworthy Learning
Federated and personalised learning, privacy-preserving AI, robust biometrics, and trustworthy generative and agentic systems.
Pai Chet Ng
Also known as Pai, Pc Ng
Assistant Professor, AI & Data Science Division, Infocomm Technology Cluster, Singapore Institute of Technology
Trustworthy AI For Multimodal Sensing In Mobile And Pervasive Systems
I joined Singapore Institute of Technology in August 2023. Previously, I was a Postdoctoral Fellow at the University of Toronto and a Research Associate at the University of Guelph. I received my PhD in Electronic and Computer Engineering from the Hong Kong University of Science and Technology.
I develop trustworthy applied AI that learns from wireless, physiological, and hyperspectral signals, connecting privacy-preserving learning with mobile devices and IoT infrastructure for healthcare, authentication, and everyday environments.
Multimodal Sensing
Wireless and physiological signals, visual biometrics, and hyperspectral measurements to understand people and their environments.
Mobile & Pervasive Systems
Intelligent sensing on mobile devices, wearables, and IoT infrastructure for accessible, deployable applications.
Selected Research
Research portfolio →
Hyperspectral imaging
Hyper-Skin
Seeing beyond RGB on consumer devices
Explore project →
Biometrics & federated learning
X-Palm & Federated-Auth
Biometrics across devices and environments
Explore project →
Trustworthy & agentic AI
ARMOR
Agentic reasoning for robust vision
Explore project →Recent Updates
Academic service →- Call for papers: TPAD@ICASSP2027. We invite you to submit papers on trustworthy perception for autonomous driving, including adversarial robustness, multimodal fusion, and foundation models.
- X-Palm was accepted to the NeurIPS 2026 Evaluations & Datasets Track.
- Chaired and organised ARTI@IEEE ICIP2026 and co-organised PFATCV@ECCV2026.
- Open-World Meme Understanding was accepted to the EMNLP 2026 main conference.
- ARMOR++ was submitted to IEEE Transactions on Reliability.
- Organised the Hyper-Object Challenge at IEEE ICASSP 2026, exploring low-cost hyperspectral imaging.
Selected Publications
Full bibliography →X-Palm: Paired Multispectral-to-Smartphone Dataset for Cross-Domain Palmprint Authentication
NeurIPS 2026, Evaluations & Datasets (E&D) Track. Accepted.
arXiv DOI: 10.48550/arXiv.2606.08437
I Know What You Meme, Even If it Emerged Today: Understanding Evolving Memes through Open-World Knowledge Acquisition
EMNLP 2026 main conference. Accepted.
arXiv DOI: 10.48550/arXiv.2606.05316
Multi-Modal Federated Learning Over Cell-Free Massive MIMO Systems for Activity Recognition
IEEE Access 13: 40844-40858 (2025).
HARWE: A multi-modal large-scale dataset for context-aware human activity recognition in smart working environments
Pattern Recognition Letters 184: 126-132 (2024).
Hyper-Skin: A Hyperspectral Dataset for Reconstructing Facial Skin-Spectra from RGB Images
NeurIPS 2023, Datasets and Benchmarks Track, pp. 24158–24170.
DOI: 10.52202/075280-1050
A Kernel Method to Nonlinear Location Estimation With RSS-Based Fingerprint
IEEE Transactions on Mobile Computing 22 (8): 4388-4404 (2023).
Stress Detection Through Wrist-Based Electrodermal Activity Monitoring and Machine Learning
IEEE Journal of Biomedical and Health Informatics 27 (5): 2155-2165 (2023).
Personal Devices for Contact Tracing: Smartphones and Wearables to Fight Covid-19
IEEE Communications Magazine 59 (9): 24-29 (2021).